Getting your Trinity Audio player ready...
|
Some European start-ups are already reportedly adopting low-cost AI models from Chinese start-up DeepSeek, in an illustration of how the Hangzhou-based company is changing expectations around AI and may force dominant companies to offer lower prices.
Hemanth Mandapati, chief executive of German start-up Novo AI, told Reuters his company initially built applications using OpenAI but switched to DeepSeek two weeks ago, with the migration taking “minutes”.
Speaking at the GoWest conference for venture capitalists in Gothenburg, Sweden, Mandapati said his company was “saving a lot of money and users don’t see any kind of a difference”.
The report, based on more than a dozen interviews with start-up executives and investors, found DeepSeek was having a significant impact on pricing and performance for AI-connected businesses, potentially pushing other AI firms to improve their models and bring down prices.
![DeepSeek artificial intelligence (AI) app displayed on a smartphone. Image credit: Unsplash 02 DeepSeek artificial intelligence (AI) app displayed on a smartphone. Image credit: Unsplash](https://www.silicon.co.uk/wp-content/uploads/2025/02/DeepSeek-artificial-intelligence-AI-app-displayed-on-a-smartphone.-Image-credit-Unsplash-02-scaled.jpg)
‘Levelling the playing field’
European companies previously struggled to adopt new technologies at the same pace as better-funded US competitors, but could now see the situation change due to DeepSeek’s influence, executives said.
Seena Rejal, chief commercial officer of UK firm NetMind.AI, another early adopter of DeepSeek, said he saw DeepSeek as a step toward democratising AI and “levelling the playing field with Big Tech”.
DeepSeek said in a technical paper that costs for training its models were a fraction of those spent on models from the likes of OpenAI’s ChatGPT or Google’s Gemini.
While DeepSeek’s training figures are being debated, the company’s financial model has allowed it to offer business pricing far below rivals.
DeepSeek’s pricing is 20 to 40 times lower than equivalent models from OpenAI, at $0.014 (1.1 pence) per 1 million input tokens for DeepSeek compared to $2.5 for OpenAI, Bernstein analysts estimated.
Fabrizio Del Maffeo, chief executive of Axelera AI, said DeepSeek showed that “bigger isn’t always better”, adding that the start-up showed it was possible to make AI more attainable and lower the total cost of ownership.
Ulrik R-T, chief executive of Denmark’s Empatik AI, said the Chinese firm was a sign that “we do not need huge budgets to be able to achieve our vision”.
Pricing pressure
DeepSeek’s influence on AI pricing may already be showing results, after Microsoft last week released OpenAI’s o1 reasoning model to all Copilot users for free, instead of restricting it to paid subscribers.
Larger companies, such as Nokia and SAP, are showing more caution, while some government bodies have warned users over DeepSeek’s security or privacy implications.
Italy’s data regulator last week suspended DeepSeek in the country over privacy issues, while the US state of Texas banned DeepSeek from government-issued devices and the US Navy told users not to use the Chinese model.
Ryan Taylor, chief revenue officer of US data analysis firm Palantir, on Monday advised clients not to use DeepSeek and said he didn’t think “any customer in the US government will be able to use it”.
Venture capitalists in the US invested some $100bn into AI start-ups last year, dominating investment, compared to about $15.8bn for Europe, according to figures from PitchBook.